Technology based on computer vision helps in choosing colors, reducing the occurrence of aesthetic problems for the patient
When a dentist defines the color of a dental prosthesis, there is a set of good practices that must be followed to ensure that the choice is as faithful as possible to the shade of natural teeth. Even so, interference from the external environment such as lighting, the specialist's level of experience and visual acuity can alter the perception of color, resulting in frustration and aesthetic problems for the patient, in addition to rework for the dentist and prosthetist, with losses. financial and waste of time.
It was with a view to eliminating or reducing the influence of these external aspects that researchers from the Computing Institute (IC-Unicamp), in partnership with Faculdade São Leopoldo Mandic, created a non-invasive technique to assist in the process of defining dental shades. The invention makes use of image analysis and computer vision techniques, an area that seeks to identify and interpret images more precisely than the human gaze, with the aim of reducing individual interference and, therefore, the degree of subjectivity in the color selection process. “Computer vision is a fantastic area, which has several very well documented and studied image processing methods and which has seen very interesting advances supported by artificial intelligence”, explains electrical engineer Willian Pulido Beneducci, who participated in the development of the Computer vision-based method for defining dental shades during his master's degree in computer science at Unicamp.
Beneducci worked on the project trying to unite the universes of knowledge in computing and dentistry. The first, led by professor Hélio Pedrini, from IC, looked for the best methods for calibrating and comparing objects in digital images. The second, guided by dentist Marcelo Lucchesi Teixeira, professor at São Leopoldo Mandic, aimed to identify the best image collection process, find volunteers and test how data collection impacts the result.
The method consists of an electronic system composed of a camera to capture the image of the tooth and software for color calibration and sample comparison. Adopting the new system has the advantage of generating photographic documentation of the patient, creating a care history and gathering information about the colors of the teeth. "The system developed, despite involving advanced computer vision techniques, is easy to use and requires few steps to automatically and objectively determine tooth shade. The solution is a fundamental step towards the use of low-cost technologies to assist professionals in the field of dentistry", explains Hélio Pedrini, professor and researcher at IC Unicamp.
Today, to use the technology, the professional needs an environment with good lighting, a common digital camera, a color calibration target, a dental color scale and a computer for imaging procedures. “In the future, however, I imagine that the camera and computer could be replaced by a smartphone with an application, as these devices have increasingly better cameras and more powerful processors”, says Beneducci.
Challenges
One of the challenges in the computing area was finding a consistent method for image color calibration, a mathematically complex process that is extremely sensitive to external noise. “It was also difficult to create a robust method for comparing target samples with the database. Due to the sensitivity of color representation, the process had to be redone and tested several times until it presented replicable results”, said Beneducci.
In dentistry, in turn, there was no public database to carry out tests. It was necessary to find volunteers in the midst of the Covid-19 pandemic and collect the images in compliance with the protection protocols recommended by the World Health Organization (WHO). It was at that moment that Lucchesi began to participate in the project, precisely to provide support at this stage.
The next steps include analyzing the feasibility of using smartphones and validating the use of the technology in a cell phone application. “The biggest takeaway from all this is that innovative research really is not trivial. We had to overcome obstacles, especially in learning processes from another area of knowledge, to have a consistent result”, concludes Beneducci.
Technology transfer
The invention was protected by the Inova Unicamp Innovation Agency with a patent deposit at the National Institute of Industrial Property (INPI). The technology is part of the University's patent and software portfolio and is available for licensing. Contact and negotiation are carried out directly with the Innovates.
This text was originally published on the Innovates Unicamp.
